Plasma appearance and distribution of E/Z and R/S isomers in plasma lipoproteins of men after single dose administration of astaxanthin

Marianne Østerlie,* Bjørn Bjerkeng,† and Synnøve Liaaen-Jensen‡

*HIST, Department of Food Science, N-7004 Trondheim, Norway, †AKVAFORSK, Institute of Aquaculture Research AS, N-6600 Sunndalsøra, Norway, and ‡Norwegian University of Science and Technology (NTNU), Organic Chemistry Laboratories, N-7491 Trondheim, Norway

Appearance, pharmacokinetics, and distribution of astaxanthin E/Z and R/S isomers in plasma and lipoprotein fractions were studied in 3 middle-aged male volunteers (37–43 years) after ingestion of a single meal containing a 100 mg dose of astaxanthin. The astaxanthin source consisted of 74% all-E-, 9% 9Z-, 17% 13Z-astaxanthin (3R,3ЈR-, 3R,3ЈS; meso-, and 3S,3ЈS-astaxanthin in a 1:2:1 ratio). The plasma astaxanthin concentration-–time curves were measured during 72 hr. Maximum levels of astaxanthin (1.3 Ϯ 0.1 mg/L) were reached 6.7 Ϯ 1.2 hr after administration, and the plasma astaxanthin elimination half-life was 21 Ϯ 11 hr. 13Z-Astaxanthin accumulated selectively, whereas the 3 and 3ЈR/S astaxanthin distribution was similar to that of the experimental meal. Astaxanthin was present mainly in very low-density lipoproteins containing chylomicrons (VLDL/CM; 36–64% of total astaxanthin), whereas low-density lipoprotein (LDL) and high-density lipoprotein (HDL) contained 29% and 24% of total astaxanthin, respectively. The astaxanthin isomer distribution in plasma, VLDL/CM, LDL, and HDL was not affected by time. The results indicate that a selective process increases the relative proportion of astaxanthin Z-isomers compared to the all-E-astaxanthin during blood uptake and that astaxanthin E/Z isomers have similar pharmacokinetics. (J. Nutr. Biochem. 11:482–490, 2000) © Elsevier Science Inc. 2000. All rights reserved.

Keywords: astaxanthin; stereoisomers; human; absorption; plasma; lipoproteins

Introduction breast milk and serum of lactating mothers, 34 , including 13 different Z-isomers have been identified.5 Dietary carotenoids from fruits and vegetables, in particular ␤ ␤ Carotenoids are absorbed from the diet by passive diffusion , -, have been associated with a decreased risk of into the intestinal mucosal cells, and transport is associated 2,3 Numer- certain cancer forms and cardiovascular diseases. with plasma lipoproteins and apparently not with specific ous carotenoids may be present in human plasma. The major 6–8 ␤ ⑀ Ј -binding proteins. Carotenoids have individ- carotenoids reported are ( , -carotene-3,3 -diol), ze- ual patterns of absorption, plasma transport, and metabolism axanthin (␤,␤-carotene-3,3Ј-diol), ␣-cryptoxanthin (␤,⑀- ␤ ␤ ␤ affected by structural differences including geometrical carotene-3-ol), -cryptoxanthin ( , -carotene-3-ol), lyco- isomerization.9 Intact carotenoids are incorporated in chy- ␺ ␺ ␤ ⑀ ␤ ␤ 4 In pene ( , -carotene), , -carotene, and , -carotene. lomicrons (CM) and transported via lymph and blood to the liver where they are partly resecreted with lipoproteins. The polar (oxygen-containing carotenoids) are dis- Preliminary results from the present study were presented at the First tributed differently among the lipoprotein fractions than are International Congress on Pigments in Food Technology, March 24–26, the non-polar .10 Carotenes are mainly carried in 1999, Seville, Spain1 and the 12th International Symposium on Carote- low-density lipoproteins (LDL; approximately 76%), noids (Book of Abstracts, p. 72), July 18–23, 1999, Cairns, Australia. whereas xanthophylls are equally distributed between LDL Address correspondence to Dr. Marianne Østerlie, HIST, Department of 11–13 Food Science, N-7004 Trondheim, Norway. and high-density lipoproteins (HDL). This may be Received February 1, 2000; accepted July 11, 2000. rationalized by a preferential solubilization of polar carote-

J. Nutr. Biochem. 11:482–490, 2000 © Elsevier Science Inc. 2000. All rights reserved. 0955-2863/00/$–see front matter 655 Avenue of the Americas, New York, NY 10010 PII S0955-2863(00)00104-2 Human uptake of astaxanthin E/Z and R/S isomers: Østerlie et al.

Figure 1 Structures of astaxanthin optical isomers all-E-(3S,3ЈS)- (1), all-E-(3R,3ЈS; meso)- (2), and all-E-(3R,3ЈR)-astaxanthin (3), and geometrical isomers all-E-(1a, 2a, 3a), 9Z-(1b, 2b,3b), 13Z-(1c, 2c, 3c), and 15Z-astaxanthin (1d, 2d, 3d). noids in the phospholipid surface and of carotenes in the widely distributed in the animal kingdom (in particular, triglyceride core.14 In humans, a considerable number of marine seafood such as salmonid fishes and shrimps) and, as studies have focused on the absorption, transport, and such, constitute part of the human diet. Carotenoids with metabolism of carotenes, whereas little information is avail- chiral centers or axes may have different optical R/S isomers able on xanthophylls.6–8 (Figure 1). E/Z-Configuration of geometrical isomers has a Astaxanthin (3,3Ј-dihydroxy-␤,␤-carotene-4,4Ј-dione) is larger influence on the overall molecular shape. The all-E-

J. Nutr. Biochem., 2000, vol. 11, October 483 Research Communication isomer of astaxanthin predominates in nature, but the 9Z-, a Beckman XL-100 ultracentrifuge equipped with a VTi 50 13Z-, and 15Z-isomers are also encountered (Figure 1). In vertical rotor (Beckman, Palo Alto, CA USA). The Quick-Seal recent years, the bioavailability of carotenoid Z-isomers has tubes were filled and unloaded by a peristaltic pump (Ismayec sa, been in focus.6–8 Astaxanthin is effective in promoting MS Reglo, Bennett & Co., N. Somerset, England). Densities of secondary immune responses against pathogens and trans- lipoprotein fractions were determined using a digital refractometer 15,16 17 (RX 5000, Atago, Tokyo, Japan). Equipment for extraction of formed cells and acts as an antioxidant. Dietary intake carotenoids included a mixer (Ultraturrax T25, Janke and Kunkel, of the astaxanthin-producing red yeast (Phaffia rhodozyma) IKA Laborteknik, Staufen, Germany) and a heating block (Reacti- or synthetic astaxanthin improves pigmentation and health therm, Pierce, Holland) equipped with a nitrogen-evaporating unit status of rainbow trout (Onchorynchus mykiss).18,19 In (Reacti-vap, Pierce, Holland). The HPLC system used was a rainbow trout, the retention of all-E-astaxanthin was much Hewlett Packard liquid chromatograph (Hewlett Packard, Palo higher than that of the astaxanthin Z-isomers.20 The higher Alto, CA USA) connected to a Hewlett Packard photodiode array absorption of all-E-astaxanthin compared to Z-astaxanthins UV-VIS detector. may be explained by the sterically more bulky Z-isomers being less permeable to the lipid membrane of the entero- Subjects and study design cytes. 13Z-astaxanthin accumulates in the liver of rainbow Three healthy male volunteers aged 37–43 years, weight 90–100 trout when compared to the all-E- and 9Z-isomers, indicat- kg (individual body-mass indexes, 27.5, 30.5, and 31.7 kg/m2, ing a selective metabolism of the E/Z-isomers.21 In a recent 22 respectively), participated in this study. Written consent was study, Elmadfa and Majchrzak failed to detect astaxan- obtained from each participant and each was instructed to abstain thin, but not , in the plasma of humans fed a from carotenoid-rich food for 3 days before and during the study. salmon meal containing canthaxanthin and astaxanthin. So A single dose of astaxanthin (100 mg, 168 ␮mol) was adminis- far, no reports on human absorption and plasma lipoprotein tered. Study subjects were instructed to eat food low in carotenoids distribution of astaxanthin E/Z or R/S isomers are available. during the 72-hr experiment. The astaxanthin dosage was prepared The objective of this study was to examine the appear- by dispersing the astaxanthin beadlets in warm water (40°C) and ance of astaxanthin E/Z or R/S isomers in the plasma of 3 mixing with olive oil (50% of meal weight) and cereals. healthy middle-aged males given a single meal containing a total of 100 mg of a mixture of all-E-, 9Z-, and 13Z- Sampling and sample preparation Ј Ј Ј astaxanthin (3R,3 R:3R,3 S:3S,3 S; ratio 1:2:1). Total astax- Blood samples (5 mL) were collected in ethylenediaminetetraace- anthin, optical isomer distribution after derivatization to the tic acid (EDTA)-containing vacuum tubes (Vacutainer, Terumo, 23 corresponding camphanates, and geometrical isomers of Belgium) equipped with a syringe after venipuncture immediately astaxanthin in the plasma were quantified by high-perfor- before and 2, 4, 6, 8, 10, 12, 24, 32, 48, and 72 hr after ingestion mance liquid chromatography (HPLC).20,24 Distribution of of the astaxanthin-containing meal. Samples were stored on ice (up E/Z-isomers in different plasma lipoprotein fractions were to 30 min) before being centrifuged (4°C, 1,500 ϫ g, 20 min). determined after ultracentrifugation in an iodixanol self- Aliquots of plasma (0.5 mL) were transferred to small test tubes generating density gradient. and stored at Ϫ80°C until carotenoids were analyzed. All samples were analyzed for astaxanthin E/Z isomer composition in intact plasma. Samples collected at 2, 4, 6, 8, 10, 12, 24, 32, and 48 hr were ultracentrifuged and pooled fractions corresponding to very Materials and methods low-density lipoprotein containing chylomicrons (VLDL/CM), Chemicals, instruments, and equipment LDL, and HDL were analyzed for astaxanthin E/Z isomer compo- sition. Samples collected 72 hr postprandially were used for Crystalline standard of all-E-astaxanthin was a gift from Hoff- analyses of astaxanthin R/S isomer composition in intact plasma. man-La Roche (Basel, Switzerland). Carophyll Pink, consisting of Astaxanthin E/Z-isomer composition in the experimental meal water-dispersible astaxanthin containing beadlets (8% astaxanthin; was determined by HPLC.20 Astaxanthin was extracted from Hoffman-La Roche, Basel, Switzerland) was used as a source of plasma and lipoprotein fractions according to Wathne et al.25 To astaxanthin, and consisted of 74% all-E-, 9% 9Z-, and 17% blood plasma samples (1.0 mL) methanol (1.0 mL; containing 500 13Z-astaxanthin (3R,3ЈR)-, (3R,3ЈS; meso)-, and (3S,3ЈS)-astaxan- ppm butylated hydroxytoluene [BHT]) was added, and the sample thin in a 1:2:1 ratio. All HPLC-grade solvents for chromatography mixed (Whirlmixer, Fisons, England) for 20 sec. CHCl3 (3 mL) and solvents for extraction (pro analysis) were purchased from was added and the sample mixed again for 20 sec. After settling for Merck (Darmstadt, Germany). Solvents for extraction were added 10 min, the sample was mixed and centrifuged (approximately 500 mg/L 2,6-di-t-butyl-p-cresol (BHT, Sigma, St. Louis, MO 1,700 g, 10 min). An aliquot (2 mL) of the hypophase containing

USA) as an antioxidant. The density gradient medium for ultra- astaxanthin dissolved in CHCl3 was pipetted into a test tube and centrifugation, iodixanol (5,5Ј-[(2-hydroxy-1-3-propanediyl)- the solvent was evaporated on a heating block (approximately Ј bis(acetylamino)] bis [N,N -bis(2,3dihydroxypropyl-2,4,6-triiodo- 40°C) using a gentle flow of nitrogen gas (N2). After evaporation, 1,3-benzenecarboxamide)], 60% w/v in water), was obtained from the sample was dissolved in 20% acetone in n-hexane (1 mL). The Nycomed Pharma AS (Optiprep, Oslo, Norway). Assay kits solution was filtered (0.45 ␮m; Minisart SRP15, Sartorius, Ger- (Cholesterol kit no. 236691, Total protein kit no. 124281, Triglyc- many) directly into the sample vials and immediately sealed. eride kit no. 701912) from Boehringer Mannheim GmbH, (Mann- Replicate analyses were performed on all samples. For determina- heim, Germany) were applied to determine the content of triglyc- tion of optical R/S isomer composition, astaxanthin was purified by erides, total protein, and cholesterol in fractions collected from thin-layer chromatography on silica gel plates (Kieselgel 60G, ultracentrifuge tubes. product no. 7731, Merck, Darmstadt, Germany). Purified astaxanthin Plasma samples were prepared from blood in a refrigerated was converted to the corresponding diesters of (-)-camphanic acid by (4°C) centrifuge (Sigma Ku¨hlzentrifuge 4kl, Osterode, Germany). reacting a dry sample of astaxanthin with (-)-camphanoyl chloride in Separation of plasma lipoproteins was performed in 15 mL dry pyridine at 0–4°C.23 Bell-top Quick-Seal tubes sealed with cordless tube toppers using Self-generating gradients containing Optiprep were used for

484 J. Nutr. Biochem., 2000, vol. 11, October Human uptake of astaxanthin E/Z and R/S isomers: Østerlie et al.

Figure 2 High-performance liquid chromatogram showing E/Z isomer composition of plasma astaxanthin 6 hr after a single oral ingestion of a meal containing 100 mg astaxanthin. Chro-

matographic conditions: H3PO4 modi- fied silica gel column,24 flow 1.0 mL/ min, detection wavelength 470 nm. Peaks were identified by retention time, co-chromatography with authentic astaxanthin isomers, and VIS spectra. Peaks 1, 2, and 3: unknown; Peak 4: all-E-astaxanthin; Peak 5: 9Z-astaxan- thin; Peak 6: 13Z-astaxanthin.

separation of blood fractions, which allowed further analysis Data analyses without removing the iodixanol.26 Polyallomer Bell-top Quick- Seal tubes (15 mL) were filled with 1.6 mL buffered saline (0.8% Pharmacokinetic parameters were estimated from plasma astaxan- NaCl in water with (n-(2-hydroxyethyl)piperazine-nЈ-(2-ethane- thin concentrations by noncompartmental methods using the PK sulfonic acid) HEPES buffer, pH 7.4) using a peristaltic pump. Solutions 2.0 program (Summit Research Services, Ashland, OH First, a layer with 6% Optiprep and thereafter, a layer with USA). The area under the plasma astaxanthin concentration–time Optiprep (1.4 mL) was mixed with plasma and buffered saline curve (AUC0–72) was calculated by using the linear trapezoidal added up to a volume of 6.7 mL (corresponding to 12% Optiprep) rule method. The AUC from the last measured time (72 hr) to was pumped underneath. A pillow (up to 0.5 mL) of 60% Optiprep infinity (AUCϱ) was calculated on the basis of AUC0–72 and an was put at the bottom of the tube to prevent contact between the extrapolated value. Oral clearance was calculated as dose/ tube wall and soluble proteins, and the tubes were ultracentrifuged AUC0–72, assuming an absorbed fraction of astaxanthin of 0.5. We ϫ used a higher absorbed fraction than indicated by the 30–40% (4 hr, 4°C, 206,000 gav) using a vertical rotor and slow 22 acceleration and deceleration below 160 rpm. Fractions of 0.5 mL absorbed amount reported by Elmadfa and Majchrzak because were unloaded starting at the bottom of the tube (highest density) the meal we used contained a high lipid content (50% olive oil). and pooled into three fractions—HDL, LDL, and VLDL/CM— Volume of distribution was estimated as oral clearance/elimination Ϯ according to contents of proteins, triglycerides, and cholesterol.27 rate. Results are presented as means SD. The densities of the fractions were determined using a refractom- Astaxanthin isomer means for individuals and for HDL, LDL, eter, and contents of triglycerides, total proteins, and cholesterol and VLDL/CM over time were compared by Student t-tests using were determined using assay kits from Boehringer Mannheim Microsoft Excel 9.0 (Microsoft Corp., Redmond, WA USA). GmbH (Mannheim, Germany). Regression analyses of astaxanthin isomer distributions versus time were performed using the REG procedure in the SAS Carotenoid analyses computer software, Version 6.12 for Windows (SAS Institute Inc., Cary, NC USA). All-E-, 9Z-, and 13Z-astaxanthin were quantified by HPLC using 24 aH3PO4 modified silica gel column, using all-E-astaxanthin (Hoffmann-La Roche, Basel, Switzerland) as an external standard. Results The flow was 1.0 mL/min and the pressure was about 22 bar. The mobile phase was renewed daily. Detection wavelength was set at The present study shows that oral administration of a 470 nm. All chromatograms were reintegrated for baseline adjust- mixture of astaxanthin E/Z isomers successfully lead to their ment. The retention times (RT) for all-E-, 9Z-, and 13Z-astaxanthin absorption into plasma without any appreciable metabolic were 10.8, 12.0, and 12.8 min, respectively. The employed transformation. Astaxanthin was not detected in the initial extinction coefficients, E1cm,1%, at 470 nm in hexane containing plasma samples. A typical chromatogram of pooled extracts 4% (v/v) CHCl were 2,100 for all-E-astaxanthin,28 1,350 and 3 of plasma illustrates the separation of astaxanthin all-E- and 1,750 for 13Z- and 9Z-astaxanthin, respectively. The E1cm,1%- values for13Z- and 9Z-astaxanthin were estimated from the HPLC Z-isomers (see Figure 2). All-E-, 9Z-, and 13Z-astaxanthin response factors of Schu¨ep and Schierle.29 Spectral fine structures were detected at 10.8, 12.0, and 12.8 min, respectively. for VIS spectra are expressed as %III/II.30 Optical isomers of Peaks 1, 2, and 3 were unidentified components with VIS ␭ ϭ astaxanthin were quantified as dicamphanates after separation by absorption characteristic for carotenoids: max 445, 470, 23 ϭ ␭ ϭ HPLC. 505 nm, %III/II 50 (possibly ), max 445 nm

J. Nutr. Biochem., 2000, vol. 11, October 485 Research Communication

Figure 3 Plasma concentration of to- tal astaxanthin (mg/L) in three middle- aged male subjects after oral ingestion of a single meal containing 100 mg astaxanthin (Carophyll Pink). Bars rep- resent standard deviation.

␭ ϭ ϭ (possibly bilirubin), and max 475 nm (possibly an concentration was found in HDL (d 1.21–1.06 g/L), less astaxanthin metabolite), respectively. Peak 3 appeared in in LDL (d ϭ 1.21–1.06 g/L) and small amounts in the chromatograms 4 hr postprandially and disappeared VLDL/CM (d ϭ1.21–1.06 g/L; Figure 4a). The highest after 10 hr. ␤,␤-Carotene elutes near the solvent front using amount of triglycerides was present in VLDL/CM (Figure our HPLC system, but was not detected in the present study. 4b), and the highest amount of cholesterol was present in Means for all plasma samples from all subjects showed a HDL (Figure 4c). distribution of astaxanthin geometrical isomers different Astaxanthin was present in all lipoprotein fractions in from the dose: 49.4 Ϯ 8.8% (range 24.2–73.1%) for varying amounts at different times post dosing (Figure 5a), all-E-astaxanthin, 13.2 Ϯ 8.8% (range 8.1–21.3%) for 9Z- although mainly in VLDL/CM (36–64%). On average, the astaxanthin, and 37.4 Ϯ 7.2% (range 18.8–54.5%) for astaxanthin content in LDL was 29.0 Ϯ 7.7% and that of 13Z-astaxanthin. The results showed a relative accumulation of HDL was 23.5 Ϯ 4.5%. The astaxanthin concentration–time the astaxanthin 13Z-isomer. Regression analysis revealed no curve of VLDL/CM exhibited two peaks: one 6–8 hr after effect of time on astaxanthin isomer distribution, except a dose ingestion, coinciding with the mean plasma astaxan- tendency (P ϭ 0.07) for the percentage of 13Z-astaxanthin to thin concentration–time curve, and one peak approximately increase. No differences were observed among individuals 16 hr after ingestion. VLDL/CM astaxanthin and triglycer- when plasma concentration time data were analyzed by Stu- ide concentrations both had maxima at 7 and 16 hr post- dent t-tests. The astaxanthin optical R/S isomer distribution in prandially (Figure 5b). In LDL, the concentration of astax- plasma closely resembled that of the dose. anthin increased until 5 hr postprandially, then decreased until 9 hr, followed by a small increase before declining. Plasma astaxanthin pharmacokinetics Regression analysis did not reveal any significant relation- ship with astaxanthin isomer distribution with time for The individual and mean astaxanthin concentrations of plasma are shown in Figure 3 and the pharmacokinetic parameters derived 72 hr post dose are presented in Table 1. Table 1 Pharmacokinetic parameters of astaxanthin after administra- Plasma astaxanthin concentrations rapidly reached a peak tion of a single oral dose of 100 mg astaxanthin (approximately 1.1 Ϯ mg/kg) to middle-aged men. An absorbed dose of 50% was assumed (Cmax) after 6.7 1.2 hr, and declined slowly thereafter. during calculations After 72 hr, 12% of Cmax (0.15 mg astaxanthin/L) still was present in the plasma. A large individual variation in Parameter Means Ϯ SD volume of distribution (range 0.2–0.6 L/kg) was due to large variations in the observed elimination half-life (t ) Ϯ 1/2 *Cmax (mg/mL) 1.3 0.1 Ϯ (range 11.4–32.1 hr) and elimination rate (range 0.022– **tmax (hr) 6.7 1.2 Elimination t (hr) 21 Ϯ 11 0.061 1/hr), not in AUC0–72 (range 34.7–42.5). 1/2 Elimination rate (1/hr) 0.042 Ϯ 0.035 ***AUC (mg hr/L) 38 Ϯ 4 Appearance of astaxanthin E/Z isomers in plasma (0–72) ***AUCϱ (mg hr/L) 42 Ϯ 3 lipoprotein fractions Oral clearance (L/hr) 0.013 Ϯ 0.01 Volume of distribution (L/kg) 0.40 Ϯ 0.2 Using a self-generating gradient of iodixanol, a density range from 1.01 to 1.21 g/L was obtained. Distributions of *Maximal concentration proteins, triglycerides, and cholesterol in a representative **Time at maximum concentration gradient are presented in Figure 4. The highest protein ***Area under the curve

486 J. Nutr. Biochem., 2000, vol. 11, October Human uptake of astaxanthin E/Z and R/S isomers: Østerlie et al.

Figure 4 Distribution of proteins, tri- glycerides, and cholesterol in high- density lipoprotein (HDL), low-density lipoprotein (LDL), and very low-density lipoprotein containing chylomicrons (VLDL/chylom.) of a representative density-separated plasma fraction us- ing a self-generating gradient of iodixa- nol. A: protein; B: triglyceride; C: cho- lesterol.

HDL, LDL, or VLDL/CM, nor did Student t-tests show any trout,21 no selectivity for astaxanthin R/S isomers was significant difference in astaxanthin isomer distribution detected in human plasma. The astaxanthin isomer distribu- between HDL, LDL, and VLDL/CM. The astaxanthin tion pattern in plasma is the combined result of possible isomers in HDL, LDL, and VLDL/CM ranged from 62.1– isomerization reactions after ingestion and isomer discrim- 66.9% for all-E-astaxanthin, 13.2–14.5% for 9Z-astaxan- ination during absorption and translocation in enterocytes thin, and 20.4–24.3% for 13Z-astaxanthin, respectively. and subsequent incorporation in blood particles. Thus, in vivo E/Z-isomerization of carotenoids during intestinal Discussion passage was recognized approximately 50 years ago31 and ␤ ␤ Astaxanthin isomers in the diet, plasma, results in a near absence of 9Z- , -carotene in postprandial plasma in humans orally administered 13C-9Z-␤,␤-caro- and lipoproteins tene;32 this is in agreement with data from other human A selective accumulation of 13Z-astaxanthin was apparent studies using large doses of equimolar mixtures of all-E- in human plasma. In line with our results with rainbow and 9Z-␤,␤-carotene.33–35 Our data were not suitable to

J. Nutr. Biochem., 2000, vol. 11, October 487 Research Communication

Figure 5 Distribution of astaxanthin and triglycerides in plasma lipoprotein frac- tions. A: distribution of astaxanthin (mg/L) in high-density lipoprotein, }; low-density lipoprotein, ■; and very low-density li- poprotein containing chylomicrons, Œ.B: distribution of astaxanthin (mg/L; Œ) and triglycerides (g/L; }) in very low-density lipoprotein containing chylomicrons. Bars represent standard deviation.

quantify the relative contribution from the different pro- anthin. Cmax was attained after about 7 hr (tmax). Reported cesses. The only similar study with accumulation of astax- values of tmax for other xanthophylls are 6 hr for 5,6-epoxy- anthin isomers is with rainbow trout in which all-E-astax- ␤,␤-carotene40,6hrfor␤-cryptoxanthin41,7,42 8,43 10,22 anthin selectively accumulated in plasma, muscle, and and 12 hr,44 respectively, for canthaxanthin, 945 and 16 hr,46 intestinal tissues, and 13Z-astaxanthin accumulated in the respectively, for lutein, 8 hr for capsanthin (3,3Ј-dihydroxy- 20,21 13 liver. However, in salmonids, tmax values for astaxan- ␤,␬-carotene-6Ј-one), and 2 and 4 hr, respectively, for the 36 37 thin have been reported in the range 18 to 30 hr, as acidic bixin and norbixin.47 These data may compared with 7 hr in the present study. These results indicate that tmax is influenced by carotenoid polarity and, clearly indicate that different species-specific mechanisms thus, gastrointestinal solubility. The acidity of bixin and govern the uptake of astaxanthin E/Z isomers in plasma. norbixin may also account for the relatively rapid blood Although no information is available on plasma, species clearance of these carotenoids.47 Absorption of carotenoids differences have previously been reported for hepatic accu- 48,49 ␤ ␤ is facilitated by dietary fat concentration and higher mulation of 9Z- and 13Z- , -carotene in human liver solubility may imply easier incorporation in gastrointestinal compared to all-E-␤,␤-carotene38 versus the reported 9Z-/ ␤ ␤ mixed micelles. The plasma elimination half-time reported all-E- , -carotene ratios of approximately 0.5 and 2.9 in rat ϭ 39 for astaxanthin (t1/2 20.8 hr) closely resembled that and chicken liver, respectively. However, liver data is not ϭ 13 reported for capsanthin (t1/2 20.1 hr), which is consid- necessarily reflecting selectivity in carotenoid absorption, ϭ 10 because complicating factors such as selective isomer me- erably less than that for lycopene (t1/2 222 hr). Thus, the tabolism may account for the observations. In conclusion, former carotenoids appear to have a more rapid metabolism our study shows that the usefulness of data on carotenoid and/or tissue deposition, as also indicated by AUCs and oral utilization in one species for comparative use in another is clearance data. Astaxanthin is virtually insoluble in water. limited. The relatively high volume of distribution therefore indicates a fairly extensive uptake, binding, or metabolism of astaxanthin Astaxanthin plasma appearance and pharmacokinetics by extravascular tissues. The different astaxanthin isomers The plasma concentration–time curve showed monophasic appeared to have similar pharmacokinetics, as judged from the kinetics of plasma appearance and disappearance of astax- regression analyses and the similar temporal distribution.

488 J. Nutr. Biochem., 2000, vol. 11, October Human uptake of astaxanthin E/Z and R/S isomers: Østerlie et al. Astaxanthin in lipoproteins References The main part of plasma astaxanthin was present in the 1 Østerlie, M., Bjerkeng, B., and Liaaen-Jensen, S. (1999). On VLDL/CM fraction (36–64% of total astaxanthin). The bioavailability and deposition of bent Z-isomers of astaxanthin. In Pigments in Food Technology (M.I.M. Mosquera, M.J. Gala´n, and remainder was nearly equally distributed between HDL and D.H. Me´ndez, eds.), pp. 157–161, Proc. 1. Int. Congr. Pigments in LDL. Our findings are in line with that for canthaxanthin, of Food Technology, March 24–26, 1999, Seville, Spain which 23.4 Ϯ 2.9% was associated with LDL 6 hr after a 2 Gaziano, J.M. and Hennekens, C.H. (1993). The role of beta- single ingestion.44 In our experiment, the time-course of carotene in the prevention of cardiovascular disease. In Carotenoids appearance of triglycerides and astaxanthin in the in Human Health, Ann. Rev. N.Y. Acad. Sci., Vol. 691 (L.M. Canfield, N.I. Krinsky, and J.A. Olson, eds.), pp. 148–155, N.Y. VLDL/CM fraction coincided (Figure 5b), as reported by Acad. Science, New York, NY, USA 44 Paetau et al. The carotenoid concentration curve in the 3 Cooper, D.A., Eldridge, R.D., and Peters, J.C. (1999). Dietary triglyceride-rich fraction after ingestion of 150 mg canthax- carotenoids and lung cancer: A review of recent research. Nutr. Rev. anthin showed three peaks at 3, 7, and 12 hr, respectively; 57, 133–145 the two earliest peaks probably resulting from variations of 4 Barua, A.B. and Furr, H.C. (1992). Extraction and analysis by 43 high-performance liquid chromatography of carotenoids in human. the speed of the lymphatic flow. The kinetics of astaxan- In Methods in Enzymology, Vol. 213 (L.B. Packer, ed.), pp. 273–281, thin in VLDL/CM and plasma are not similar, explained by Academic Press, San Diego, CA, USA the fact that plasma responses do not allow distinction 5 Khachik, F., Spangler, C.J., Smith, J.C., Canfield, L.M., Steck, A., between newly absorbed carotenoids and those mobilized and Pfander, H. (1997). Identification, quantification, and relative concentrations of carotenoids and their metabolites in human milk from endogenous sources. The observed kinetics of astax- and serum. Anal. Chem. 69, 1873–1881 anthin in lipoprotein fractions and the rapid increase of 6 Olson, J.A. (1994). Absorption, transport, and metabolism of caro- astaxanthin content in both VLDL/CM and LDL are similar tenoids in humans. Pure Appl. Chem. 66, 1011–1016 to plasma appearance of canthaxanthin for 12 hr44 and 7 Parker, R.S. (1996). Absorption, metabolism, and transport of 13 carotenoids. FASEB J. 10, 542–551 capsanthin after a single ingestion of paprika juice. A 8 Furr, H.C. and Clark, R.M. (1997). Intestinal absorption and tissue preferential uptake of the xanthophylls lutein and zeaxan- distribution of carotenoids. J. Nutr. Biochem. 8, 364–377 thin, as compared with that of all-E-␤,␤-carotene in CM, 9 Clevidence, B.A. and Bierer, J.G. (1993). Association of carotenoids was shown after Betatene ingestion, with maximum uptake with human plasma lipoproteins. In Methods in Enzymology, Vol. at 9 hr and distinct decline from 9 to 12 hr.45 In another in 214 (L.B. Packer, ed.), pp. 33–46, Academic Press, San Diego, CA, USA vitro study modeling biological emulsions, more polar 10 Romanchik, J.E., Morel, D.W., and Harrison, E.H. (1995). Distribu- carotenoids such as astaxanthin were solubilized in the tions of carotenoids and ␣-tocopherol among lipoproteins do not phospholipid surface of lipid droplets, and the apolar change when human plasma is incubated in vitro. J. Nutr. 125, carotenoids were associated with the triglyceride core.14 2610–2617 11 Krinsky, N.I., Cornwell, D.G., and Oncley, J.L. (1958). The trans- The transfer of the polar carotenoids between bioemulsion- port of vitamin A and carotenoids in human plasma. Arch. Biochem. like lipoproteins was easier than that of the apolar carote- Biophys. 73, 233–246 noids, which required triglyceride lipolysis by lipase. The 12 Goulinet, S. and Chapman M. J. (1997). Plasma LDL and HDL similar distribution of the different astaxanthin E/Z isomers subspecies are heterogeneous in particle content of tocopherols and in the plasma protein fractions during the trial may indicate oxygenated and hydrocarbon carotenoids. Arterioscler. Thromb. Vasc. Biol. 17, 786–796 that isomer selection is associated with the digestive pro- 13 Oshima, S., Sakamoto, H., Ishiguro, Y., and Terao, J. (1997). cess, although isomerization may also occur. Analysis of Accumulation and clearance of capsanthin in blood plasma after the blood of individuals given doses of purified astaxanthin E/Z ingestion of paprika juice in men. J. Nutr. 127, 1475–1479 isomers is required to obtain more information on isomer- 14 Borel, P., Grolier, P., Armand, M., Partier, A., Lafont, H., Lairon, D., and Azais-Braesco, V. (1996). Carotenoids in biological emulsions: ization during digestion. Solubility, surfaces-to-core distribution, and release from lipid drop- lets. J. Lipid Res. 37, 250–261 15 Jyonouchi, H., Sun, S., Tomita, Y., and Gross, M.D. (1995). Astaxanthin, a carotenoid without vitamin A activity, augments Conclusion antibody responses in cultures including T-helper cell clones and suboptimal doses of antigen. J. Nutr. 125, 2483–2492 As reported for other xanthophylls cited in this article, 16 Bertram, J.S. (1999). Carotenoids and gene regulation. Nutr. Rev. 57, dietary astaxanthin was readily absorbed and incorporated 182–191 in human plasma lipoproteins at a considerable degree. 17 Martin, H.-D., Ja¨ger, C., Ruck, C., Schmidt, M., Walsh, R., and Compared to rainbow trout,20,21 different E/Z isomers were Paust, J. (1999). Anti- and prooxidant properties of carotenoids. J. Prakt. Chem. 341, 302–308 selectively accumulated in human plasma, indicating differ- 18 Nakano, T., Tosa, M., and Takeuchi, M. (1995). Improvement of ent uptake mechanisms. The bioavailability of carotenoids biochemical features in fish health by red yeast and synthetic is influenced by a number of factors, termed SLAMENGHI astaxanthin. J. Agric. Food Chem. 43, 1570–1573 (Species of carotenoids, molecular Linkage, Amount of 19 Nakano, T., Kanmuri, T., Sato, M., and Takeuchi, M. (1999). Effect carotenoids consumed in a meal, Matrix, Effectors of of astaxanthin rich red yeast (Phaffia rhodozyma) on oxidative stress in rainbow trout. Biochim. Biophys. Acta 1426, 119–125 absorption and bioconversion, Nutrient status of the host, 20 Bjerkeng, B., Følling, M., Lagocki, S., Storebakken, T., Olli, J.J., Genetic factors, Host-related factors, and mathematical and Alsted, N. (1997). Bioavailability of all-E-astaxanthin and Interactions), as reviewed by Castenmiller and West.50 Z-isomers of astaxanthin in rainbow trout (Oncorhynchus mykiss). Apparently, none of these factors are able to explain Aquaculture 157, 63–82 21 Østerlie, M., Bjerkeng, B., and Liaaen-Jensen, S. (1999). Accumu- differences in uptake between species. Future studies should lation of all-E,9Z and 13Z geometrical isomers and 3 and 3Ј RS seek to establish the biochemical basis for species differ- optical isomers of astaxanthin in rainbow trout (Oncorhynchus ences in carotenoid utilization. mykiss) is selective. J. Nutr. 129, 391–398

J. Nutr. Biochem., 2000, vol. 11, October 489 Research Communication

22 Elmadfa, I. and Majchrzak, D. (1999). Absorption und transport von 36 Gobantes, I., Choubert, G., Laurentie, M., Milicua, J.-C.G., and astaxanthin und canthaxanthin beim menschen nach einer lachsmahl- Gomez, R. (1997). Astaxanthin and canthaxanthin kinetics after zeit. Erna¨hr.-Umschau 46, 173–177 ingestion of individual doses by immature rainbow trout Oncorhyn- 23 Vecchi, M. and Mu¨ller, R.K. (1979). Separation of (3S,3ЈS)-, chus mykiss. J. Agric. Food Chem. 45, 454–458 (3R,3ЈR)- and (3S,3ЈR)- astaxanthin via (-)-camphanic acid esters. J. 37 Aas, G.H., Bjerkeng, B., Storebakken, T., and Ruyter, B. (1999). High Resolution Chromatogr. Chromatogr. Commun. 2, 195–196 Blood appearance, metabolic transformation and plasma transport 24 Vecchi, M., Glinz, E., Meduna, V., and Schiedt, K. (1987). HPLC proteins of 14C-astaxanthin in Atlantic salmon (Salmo salar L.). Fish separation and determination of astacene, semiastacene, astaxanthin, Physiol. Biochem. 21, 325–334 and other keto-carotenoids. J. High Resolution Chromatogr. Chro- 38 Stahl, W., Schwarz, W., Sundquist, A.R., and Sies, H. (1992). matogr. Commun. 10, 348–351 Cis-trans isomers of lycopene and ␤-carotene in human serum and 25 Wathne, E., Bjerkeng, B., Storebakken, T., Vassvik, V., and Odland, tissues. Arch. Biochim. Biophys. 294, 173–177 A.B. (1998). Pigmentation of Atlantic salmon (Salmo salar) fed 39 Ben-Amotz, A., Mokady, S., Edelstein, S., and Avron, M. (1989). astaxanthin in all meals or in alternating meals. Aquaculture 159, Bioavailability of a natural isomer mixture as compared with 217–231 synthetic all-trans-␤-carotene in rats and chicks. J. Nutr. 119, 26 Graham, J.M., Higgins, J.A., Gillot, T., Taylor, T., Wilkinson, J., 1013–1019 Ford, T., and Billington, D. (1996). A novel method for the rapid 40 Barua, A.B. (1999). Intestinal absorption of epoxy-␤-carotenes by separation of plasma lipoproteins using self-generating gradients of humans. Biochem. J. 339, 359–362 iodixanol. Atherosclerosis 124, 125–135 41 Wingerat, T., Stahl, W., and Sies, H. (1998). ␤-Cryptoxanthin 27 Linscheer, W.G. and Vergroesen, A.J. (1994). Lipids. In Modern selectively increases in human chylomicrons upon ingestion of Nutrition in Health and Disease, 8th ed. (M.E. Shils, J.A. Olson, and tangerine concentrate rich in ␤-cryptoxanthin esters. Arch. Biochem. M. Shike, eds.), pp. 47–88, Lea & Febiger, Philadelphia, PA, USA Biophys. 324, 385–390 28 Britton, G. (1995). UV/VIS spectroscopy. In Carotenoids, Vol. 1B, 42 White, W.S., Stacewicz-Sapuntzakis, M., Erdman, J.W., and Bowen, Spectroscopy (G. Britton, S. Liaaen-Jensen, and H. Pfander, eds.), P.E. (1994). Pharmacokinetics of ␤-carotene and canthaxanthin after pp. 13–62, Birkha¨user, Basel, Switzerland ingestion of individual and combined doses by human subjects. 29 Schu¨ep, W. and Schierle, J. (1995). Example 9: Astaxanthin. In J. Am. Coll. Nutr. 13, 665–671 Carotenoids, Vol. 1A, Isolation and Analysis (G. Britton, S. Liaaen- 43 Ku¨bler, W. (1989). Pharmacokinetic implications of single and Jensen, and H. Pfander, eds.), pp. 273–276, Birkha¨user, Basel, repeated dosage. Int. J. Vit. Nutr. Res. 30(suppl), S25–S34 Switzerland 44 Paetau, I., Chen, H., Goh, N.M.-Y., and White, W.S. (1997). 30 Ke, B., Imsgard, F., Køsen, H., and Liaaen-Jensen, S. (1970). ␤ Electronic spectra of carotenoids at 77°K. Biochim. Biophys. Acta Interactions in the postprandial appearance of -carotene and can- 210, 139–152 thaxanthin in plasma triacylglycerol-rich lipoproteins in humans. 31 Kemmerer, A.R. and Fraps, G.S. (1945). The vitamin A activity of Am. J. Clin. Nutr. 66, 1133–1143 neo-␤-carotene U and its steric rearrangement in the digestive tract 45 Ga¨rtner, C., Stahl, W., and Sies, H. (1996). Preferential increase in of rats. J. Biol. Chem. 161, 305–309 chylomicron levels of xanthophylls lutein and compared ␤ 32 You, C.-S., Parker, R.S., Goodman, K.J., Swanson, J.E., and Corso, to -carotene in the human. Int. J. Vitamin Nutr. Res. 66, 119–125 T.N. (1996). Evidence of cis-trans isomerization of 9-cis-␤-carotene 46 Kostic, D., White, W.S., and Olson, J.A. (1995). Intestinal absorp- ␤ during absorption in humans. Am. J. Clin. Nutr. 64, 177–183 tion, serum clearance and interactions between lutein and -carotene 33 Stahl, W., Schwarz, W., and Sies, H. (1993). Human serum concen- when administered to human adults in separate or combined oral trations of all-trans-␤- and ␣-carotene but not 9-cis-␤-carotene doses. Am. J. Clin. Nutr. 62, 604–610 increase upon ingestion of a natural isomer mixture obtained from 47 Levy, L.W., Regalado, E., Navarrete, S., and Watkins, R.H. (1997). Dunaliella salina (Betatene). J. Nutr. 123, 847–851 Bixin and norbixin in human plasma: Determination and study of the 34 Stahl, W., Schwarz, W., van Laar, J., and Sies, H. (1995). All-trans- absorption of a single dose of annatto food color. Analyst 122, ␤-carotene preferentially accumulates in human chylomicron and 977–980 very low density lipoproteins compared with 9-cis geometrical 48 Dimitrov, N.V., Meyers, C., Ullrey, D.E., Chenoweth, W., Mich- isomer. J. Nutr. 125, 2128–2133 elakis, A., Malone, W., Boone, C., and Fink, G. (1988). Bioavail- 35 Doering, W.E., von Sotiriou-Leventis, C., and Roth, W.R. (1995). ability of ␤-carotene in humans. Am. J. Clin. Nutr. 48, 298–304 Thermal interconversions among 15-cis-, 13-cis- and all-trans-␤- 49 Prince, M.R. and Frisoli, J.K. (1993). Beta-carotene accumulation in carotene: Kinetics, Arrhenius parameters, thermochemistry, and serum and skin. Am. J. Clin. Nutr. 57, 175–181 potential relevance to anticarcinogenicity of all-trans-␤-carotene. 50 Castenmiller, J.J.M. and West, C.E. (1998). Bioavailability and J. Am. Chem. Soc. 117, 2747–2757 bioconversion of carotenoids. Ann. Rev. Nutr. 18, 19–38

490 J. Nutr. Biochem., 2000, vol. 11, October